From the review: "This tactic, in which A.I. companies promise wild new features and deliver a half-baked product, is becoming a trend that is bound to confuse and frustrate people."
I wouldn't necessarily call GPT-4o half-baked - it's feels like a slight improvement over GPT-4 Turbo to me (albeit with possibly a higher hallucination rate, at least it seems to me). But much of what was demoed is not yet available, and some of the use cases that were shown and are supported can have really variable results. Example: math tutoring is one that I would approach with extreme caution, at least for now.
The race by companies to promote generative AI, including in use cases it does not (yet) perform that well in, is having I think a large negative impact on trust and perception of AI overall. I include Google's generative AI summaries applied to search in that regard. With search results being as gunked-up as they are with spam and other attempts to game the SEO system, it's no wonder we are seeing the sort of AI summaries created that have made the news.
![View profile for Steven Vonder Haar, graphic]()
So often, it seems, in the AI world we seem to have an abundance of application demos that do not yet translate into usable solutions in the real-world. Increasingly, the world seems to be taking notice.
Note the level of frustration conveyed in this review from The New York Times of OpenAI's ChatGPT-4o, which fails to deliver on several aspects of this chatbot iteration described at its launch.
The flimsy, mercurial performance of solutions that fall short of their promise is giving AI a public relations black eye. Fundamental change still is coming to the digital marketplace. The exponential increases in computing power embodied by AI will drive fundamental change in how software is developed and what software applications are able to accomplish.
When this transformation becomes a reality, however, we may not be crediting "artificial intelligence." The term is nearing the point of becoming "tech toxic" - more representative of failed promises than the awesome possibilities it should be associated with.
#ArtificialIntelligence #TechHype #BrokenPromises
![The New ChatGPT Offers a Lesson in A.I. Hype]()
The New ChatGPT Offers a Lesson in A.I. Hype
https://www.nytimes.com